An Engineering Approach to Computer Networking Naming and Addressing An Engineering Approach to Computer Networking

Outline Names and addresses Hierarchical naming Addressing Addressing in the telephone network Addressing in the Internet ATM addresses Name resolution Finding datalink layer addresses

Names and addresses Names and addresses both uniquely identify a host (or an interface on the host) %nslookup Default Server: DUSK.CS.CORNELL.EDU Address: 128.84.227.13 > underarm.com Name: underarm.com Address: 206.128.187.146 Resolution: the process of determining an address from a name

Why do we need both? Names are long and human understandable wastes space to carry them in packet headers hard to parse Addresses are shorter and machine understandable if fixed size, easy to carry in headers and parse Indirection multiple names may point to same address can move a machine and just update the resolution table

Hierarchical naming Goal: give a globally unique name to each host Naïve approach: ask other naming authorities before choosing a name doesn’t scale (why?) not robust to network partitions Instead carve up name space (the set of all possible names) into mutually exclusive portions => hierarchy

Hierarchy A wonderful thing! scales arbitrarily guarantees uniqueness easy to understand Example: Internet names use Domain name system (DNS) global authority (Network Solutions Inc.) assigns top level domains to naming authorities (e.g. .edu, .net, .cz etc.) naming authorities further carve up their space all names in the same domain share a unique suffix

Addressing Addresses need to be globally unique, so they are also hierarchical Another reason for hierarchy: aggregation reduces size of routing tables at the expense of longer routes

Addressing in the telephone network Telephone network has only addresses and no names (why?) E.164 specifications ITU assigns each country a unique country code Naming authority in each country chooses unique area or city prefixes Telephone numbers are variable length this is OK since they are only used in call establishment Optimization to help dialing: reserve part of the lower level name space to address top level domains e.g. in US, no area code starts with 011, so 011 => international call => all other calls need fewer digits dialed

Addressing in the Internet Every host interface has its own IP address Routers have multiple interfaces, each with its own IP address Current version of IP is version 4, addresses are IPv4 addresses 4 bytes long, two part hierarchy network number and host number boundary identified with a subnet mask can aggregate addresses within subnets

Address classes First cut fixed network-host partition, with 8 bits of network number too few networks! Generalization Class A addresses have 8 bits of network number Class B addresses have 16 bits of network number Class C addresses have 24 bits of network number Distinguished by leading bits of address leading 0 => class A (first byte < 128) leading 10 => class B (first byte in the range 128-191) leading 110 => class C (first byte in the range 192-223)

Address evolution This scheme was too inflexible Three extensions subnetting CIDR dynamic host configuration

Subnetting Allows administrator to cluster IP addresses within its network

CIDR Scheme forced medium sized nets to choose class B addresses, which wasted space Address space exhaustion Solution allow ways to represent a set of class C addresses as a block, so that class C space can be used use a CIDR mask idea is very similar to subnet masks, except that all routers must agree to use it subnet masks are not visible outside the network (why?)

CIDR (contd.)

Dynamic host configuration Allows a set of hosts to share a pool of IP addresses Dynamic Host Configuration Protocol (DHCP) Newly booted computer broadcasts discover to subnet DHCP servers reply with offers of IP addresses Host picks one and broadcasts a request to a particular server All other servers withdraw offers, and selected server sends an ack When done, host sends a release IP address has a lease which limits time it is valid Server reuses IP addresses if their lease is over Similar technique used in Point-to-point protocol (PPP)

IPv6 32-bit address space is likely to eventually run out IPv6 extends size to 128 bits Main features classless addresses multiple levels of aggregation are possible registry provider subscriber subnet several flavors of multicast anycast interoperability with IPv4

ATM network addressing Uses Network Service Access Point (NSAP) addresses Variable length (7-20 bytes) Several levels of hierarchy national or international naming authority addressing domain subnet

Name resolution Done by name servers essentially look up a name and return an address Centralized design consistent single point of failure concentrates load

DNS Distributed name server A name server is responsible (an authoritative server) for a set of domains May delegate responsibility for part of a domain to a child Root servers are replicated If local server cannot answer a query, it asks root, which delegates reply Reply is cached and timed out

Finding datalink layer addresses Datalink layer address: most common format is IEEE 802 Need to know datalink layer address typically for the last hop

ARP To get datalink layer address of a machine on the local subnet Broadcast a query with IP address onto local LAN Host that owns that address (or proxy) replies with address All hosts are required to listen for ARP requests and reply including laser printers! Reply stored in an ARP cache and timed out In point-to-point LANs, need an ARP server register translation with server ask ARP server instead of broadcasting